09-28-2014 03:45 AM
I have a data set with a variable marked as Target Variable for E-Miner. It is evenly split into Very_Bad & Other. All I am trying to do is to split this dataset into Train, Validate, & Test using Data partition node. When i use the node, the data is partitioned solely for 1 target class, leaving the other untouched. So my datasets after the Data partition node Train, Validate, & Test now solely contain only 1 Target Class (Very_bad)....Why ? They should contain a decent mixture of both target classes. Why is it happening. Getting frustrated ?